6 The Russian school Gregory Falkovich The towering figure of Kolmogorov and his very productive school is what was perceived in the twentieth century as the Russian school of turbulence. How- ever, important Russian contributions neither start nor end with that school. 6.1 Physicist and pilot ...thebombswerefalling almost the way the theory predicts. To have conclusive proof of the theory I’m going to fly again in a few days. A.A. Friedman, letter to V.A. Steklov, 1915 What seems to be the first major Russian contribution to the turbulence the- ory was made by Alexander Alexandrovich Friedman, famous for his work on non-stationary relativistic cosmology, which has revolutionized our view of the Universe. Friedman’s biography reads like an adventure novel. Alexan- der Friedman was born in 1888 to a well-known St. Petersburg artistic family (Frenkel, 1988). His father, a ballet dancer and a composer, descended from a baptized Jew who had been given full civil rights after serving 25 years in the army (a so-called cantonist). His mother, also a conservatory graduate, was a daughter of the conductor of the Royal Mariinsky Theater. His parents di- vorced in 1897, their son staying with the father and becoming reconciled with his mother only after the 1917 revolution. While attending St. Petersburg’s second gymnasium (the oldest in the city) Friedman befriended a fellow stu- dent Yakov Tamarkin, who later became a famous American mathematician and with whom he wrote their first scientific works (on number theory, re- ceived positively by David Hilbert). In 1906, Friedman and Tamarkin were admitted to the mathematical section of the Department of Physics and Math- ematics of Petersburg University where they were strongly influenced by the 209 Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 23 Jul 2018 at 14:46:21, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139018241.007 210 Falkovich great mathematician V.A. Steklov who taught them partial differential equa- tions and regularly invited them to his home (with another fellow student V.I. Smirnov who later wrote the well-known Course of Mathematics, the first volume with Tamarkin). As his second, informal, teacher Alexander always mentioned Paul Ehrenfest who was in St. Petersburg in 1907–1912 and later corresponded with Friedman. Friedman and Tamarkin were among the few mathematicians invited to attend the regular seminar on theoretical physics in Ehrenfest’s apartment. Apparently, Ehrenfest triggered Friedman’s interest in physics and relativity, at first special and then general. During his grad- uate studies, Alexander Friedman worked on different mathematical subjects related to a wide set of natural and practical phenomena (among them on poten- tial flow, corresponding with Joukovsky, who was in Moscow). Yet after get- ting his MSc degree, Alexander Friedman was firmly set to work on hydrody- namics and found employment in the Central Geophysical Laboratory. There, the former pure mathematician turned into a physicist, not only doing theory but also eagerly participating in atmospheric experiments, setting the measure- ments and flying on balloons. It is then less surprising to find Friedman flying a plane during World War I, when he was three times decorated for bravery. He flew bombing and reconnaissance raids, calculated the first bombardment tables, organized the first Russian air reconnaissance service and the factory of navigational devices (in Moscow, with Joukovsky’s support), all the while publishing scientific papers on hydrodynamics and atmospheric physics. Af- ter the war ended in 1918, Alexander Alexandrovich was given a professorial position at Perm University (established in 1916 as a branch of St. Petersburg University), which boasted at that time Tamarkin, Besikovich and Vinogradov among the faculty. In 1920 Friedman returned to St. Petersburg. Steklov got him a junior position at the University (where George Gamov learnt relativity from him). Soon Friedman was teaching in the Polytechnic as well, where L.G. Loitsyansky was one of his students. In 1922 Friedman published his famous work On the curvature of space where the non-stationary Universe was born (Friedman, 1922). The conceptual novelty of this work is that it posed the task of describing the evolution of the Universe, not only its structure. The next year saw the dramatic exchange with Einstein, who at first published the paper that claimed that Friedman’s work contained an error. Instead of public polemics, Friedman sent a personal letter to Einstein where he elaborated on the details of his derivations. After that, Einstein published the second paper admitting that the error was his. In 1924 Friedman published his work, described below, that laid down the foundations of the statistical theory of turbulence structure. In 1925 he made a record-breaking balloon flight to the height of 7400 meters to study atmospheric vortices and make medical self-observations. His personal Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 23 Jul 2018 at 14:46:21, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139018241.007 6: The Russian school 211 life was quite turbulent at that time too: he was tearing himself between two women, a devoted wife since 1913 and another one pregnant with his child (“I do not have enough willpower at the moment to commit suicide” he wrote in a letter to the mother of his future son). On his way back from summer vacations by train in the Crimea, Alexander Friedman bought a nice-looking pear at a Ukrainian train station, did not wash it before eating and died from typhus two weeks later. Friedman’s work on turbulence theory was done in conjunction with his stu- dent Keller and was based on the works of Reynolds and Richardson, both cited extensively in Friedman and Keller (1925). Recall that Richardson de- rived the equations for the mean values which contained the averages of non- linear terms that characterize turbulent fluctuations. Friedman and Keller cite Richardson’s remark that such averaging would work only in the case of a so-called time separation when fast irregular motions are imposed on a slow- changing flow, so that the temporal window of averaging is in between the fast and slow timescales. For the first time, they then formulated the goal of writing down a closed set of equations for which an initial value problem for turbulent flow can be posed and solved. The evolutionary (then revolutionary) approach of Friedman to the description of the small-scale structure of turbulence paral- lels his approach to the description of the large-scale structure of the Universe. Achieving closure in the description of turbulence is nontrivial since the hy- drodynamic equations are nonlinear. Indeed, if v is the velocity of the fluid, then Newton’s second law gives the acceleration of the fluid particle: dv ∂v = + (v∇)v = force per unit mass. (6.1) dt ∂t Whatever the forces, the acceleration already contains the second (inertial) term, which makes the equation nonlinear. Averaging the fluid dynamical equa- tions, one expresses the time derivative of the mean velocity, ∂ v /∂t, via the quadratic mean (v∇)v . Friedman and Keller realized that meaningful closure can only be achieved by introducing correlation functions between different points in space and different moments in time. Their approach was intended for the description of turbulence superimposed on a non-uniform mean flow. Writing the equation for the two-point function ∂ v1v2 /∂t, they then derived the closed system of equations by decoupling the third moment via the second i j k = + ··· moment and the mean: v1v2v2 v1i v2 jv2k (Friedman and Keller, 1925). It is interesting that Friedman called the correlation functions “moments of conservation” (Erhaltungsmomenten) as they express “the tendency to pre- serve deviations from the mean values” in a curious resemblance to the modern approach based on martingales or zero modes. The work was presented at the Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 23 Jul 2018 at 14:46:21, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139018241.007 212 Falkovich First International Congress on Applied Mechanics in Delft in 1925. During the discussion after Friedman’s talk he made it clear that he was aware that the approximation is crude and that time averages are not well-defined. He stressed that his goal was pragmatic (predictive meteorology) and that only a consistent theory of turbulence can pave the way for dynamical meteorology: “Instruments give us mean values while hydrodynamic equations are applied to the values at a given moment”. The introduction of correlation functions was thus the main contribution to turbulence theory made by Alexander Friedman, a great physicist and a pilot. One year after Friedman’s death, the seminal paper of Richardson on atmo- spheric diffusion appeared. I cannot resist imagining what would have hap- pened if Friedman saw this paper and made a natural next step: to incorpo- rate the idea of cascade and the scaling law of Richardson’s diffusion into the Friedman–Keller formalism of correlation functions and to realize that the third moment of velocity fluctuations, which they neglected, is crucial for the description of the turbulence structure. As it happened, this was done 15 years later by another great Russian scientist, mathematician Andrei Nikolaevich Kolmogorov. 6.2 Mathematician At any moment, there exists a narrow layer between trivial and impossible where mathematical discoveries are made. Therefore, an applied problem is either solved trivially or not solved at all.